This paper first reviews previous relevant efforts to bring energy system models and life-cycle assessments together. In section 3, we set out the method we have used, including the details of the EEIO model, the TIMES model, the procedures for linking them, and the scenarios examined. Section 4 then provides results relating to each of the research questions, while section 5 draws key conclusions. The paper is accompanied by a supplementary file that includes considerable further detail on the modelling approach, the data used, and the results. 2 Literature review Bottom-up ESOMs (such as MARKAL [Loulou et al. 2004] and TIMES [Loulou et al. 2004]) and MESSAGE (Messner and Schrattenholzer 2000)) provide a detailed depiction of the energy system, with explicit representation of primary extraction of energy resources, processing and conversion, delivery to consumers, and end-use (DeCarolis et al. 2017). Such models account for some emissions associated with upstream extraction (flaring, for example), and they account for the efficiency losses and energy inputs associated with conversion and processing (e.g. in refineries and power stations), with transmission and distribution (losses in electricity transmission lines, energy use in fuel distribution, etc.), and efficiency losses in end use devices. Many sources of fossil fuel chain indirect emissions are thus already included in the default setup of TIMES (see the supplementary file for further details). ESOMs are "demand-driven" in the sense that the energy service demands across the economy are a key exogenous input into the models. Energy service demands associated with